What is Active / Active Array ? - Storage

This is a discussion on What is Active / Active Array ? - Storage ; nope, powerpath is not influenced in anyway by Navisphere Agent "Peter" wrote in message news:eAy4281HHHA.3952@TK2MSFTNGP02.phx.gbl... > Besides, I would like to know whether Navisphere Agent affects PowerPath > Failover software ? > > Peter > > "Edwin vMierlo" wrote in ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 22 of 22

Thread: What is Active / Active Array ?

  1. Re: What is Active / Active Array ?

    nope, powerpath is not influenced in anyway by Navisphere Agent


    "Peter" wrote in message
    news:eAy4281HHHA.3952@TK2MSFTNGP02.phx.gbl...
    > Besides, I would like to know whether Navisphere Agent affects PowerPath
    > Failover software ?
    >
    > Peter
    >
    > "Edwin vMierlo" wrote in message
    > news:eh3DV8pHHHA.1468@TK2MSFTNGP04.phx.gbl...
    > > They can talk to you the security setup as well (I guess, we only have

    EMC
    > > kit directly from EMC, I have no explicit experience dealing with Dell

    in
    > > this matter)
    > >
    > > Get the connectivity guide (from EMC)... ask the Dell boys to give you a
    > > (soft-)copy, it is a handy document explaining setups in general. You

    can
    > > base your internal setup and maintenance procedures on this doc.
    > >
    > > good luck,
    > > Edwin.
    > >
    > >
    > > "Peter" wrote in message
    > > news:urd80xpHHHA.3780@TK2MSFTNGP02.phx.gbl...
    > >> Yes, the host can access the disks.
    > >>
    > >> Since we purchased the array from DELL, we are not able to talk to EMC
    > >> support. We have to talk to DELL staff.
    > >>
    > >> Regards,
    > >> Peter
    > >>
    > >> "Edwin vMierlo" wrote in

    message
    > >> news:OCcvyrpHHHA.420@TK2MSFTNGP06.phx.gbl...
    > >> > Peter,
    > >> >
    > >> > non-functioning of Navisphere Agent will not limit you accessing your
    > >> > disks,
    > >> > if your HBA's from the host are properly logged in and registered,

    and
    > > in
    > >> > the proper storage group, you host will be able access the disk.
    > >> >
    > >> > Q: can you access the disks from your host ?
    > >> >
    > >> > then in regards to non-functioning of the Agent, there are some
    > >> > security
    > >> > settings which need to match between host and array. Easiest is to

    open
    > > a
    > >> > case with EMC, they can talk you through this is 5 minutes.

    > > Alternatively,
    > >> > there is a connectivity guide on the EMC website which will detail

    all
    > >> > steps
    > >> > needed. I would like to talk you through it by newsgroup posts, but

    the
    > >> > procedure has too many variables to be 100% accurate.
    > >> >
    > >> >
    > >> >
    > >> > "Peter" wrote in message
    > >> > news:uJhuY5oHHHA.1264@TK2MSFTNGP03.phx.gbl...
    > >> >> Dear Edwin and JM,
    > >> >>
    > >> >> DELL Engineer has updated the SAN Firmware and Hosts HBA /

    Navisphere
    > > and
    > >> >> Powerpath for us during the weekend.
    > >> >>
    > >> >> However, from Navisphere web interface, we find that the Navisphere

    > > Agent
    > >> > in
    > >> >> one of the host (Windows 2003 Server) cannot be retrieved by the
    > >> >> Array.
    > >> >> I
    > >> >> would like to what will be the consequence ? Is it a serious

    problem
    > > or
    > >> >> just the drives information is not presented ?
    > >> >>
    > >> >> I have checked Powerpath and it seems to be OK.
    > >> >>
    > >> >> Regards,
    > >> >> Peter
    > >> >> "Edwin vMierlo" wrote in

    > > message
    > >> >> news:ONtpYYeGHHA.3952@TK2MSFTNGP02.phx.gbl...
    > >> >> > response in-line
    > >> >> >
    > >> >> >
    > >> >> > "Peter" wrote in message
    > >> >> > news:ObW4iHeGHHA.924@TK2MSFTNGP02.phx.gbl...
    > >> >> >> Dear JM and Edwin,
    > >> >> >>
    > >> >> >> Thank you for your advice.
    > >> >> >> Since we are using FC disk and not ATA one, the consultant's

    advice
    > >> > looks
    > >> >> >> not valid.
    > >> >> >>
    > >> >> >> Does ATA disk means SATA disk ?
    > >> >> >> Can we use FC disk with different speed ?
    > >> >> >
    > >> >> > This might be depending on the Flare-code revision, and other
    > >> >> > things,
    > >> > you
    > >> >> > really need to ask that question to your EMC rep.
    > >> >> >
    > >> >> >> This is because from EMC web
    > >> >> >> site, I find that they support SAN disk with higher rpm (Of

    course,
    > >> > 300GB
    > >> >> > HD
    > >> >> >> is still not available).
    > >> >> >>
    > >> >> >> The consultant also suggests us to upgrade to the latest firmware

    > > (The
    > >> >> >> one
    > >> >> >> we are using is more than 2 years old). Is this suggestion valid

    ?
    > >> >> >
    > >> >> > Most of the time; yes. Keep abrest with latest versions of

    software
    > > and
    > >> >> > firmware. BUT If you have a test or try-out system, I would highly
    > >> >> > recommend
    > >> >> > to run upgrades on test systems first.
    > >> >> >
    > >> >> >>
    > >> >> >> Regards,
    > >> >> >> Peter
    > >> >> >>
    > >> >> >> "Edwin vMierlo" wrote in
    > >> > message
    > >> >> >> news:eup4p1cGHHA.1240@TK2MSFTNGP03.phx.gbl...
    > >> >> >> > That could explain it, I have no experience with CX series with

    > > ATA
    > >> >> >> > drives,
    > >> >> >> > only FC drives.
    > >> >> >> >
    > >> >> >> > thanks for explaingin JM !
    > >> >> >> > rgds,
    > >> >> >> > Edwin.
    > >> >> >> >
    > >> >> >> >
    > >> >> >> > "JM" wrote in message
    > >> >> >> > news:%23jOAPHVGHHA.1280@TK2MSFTNGP04.phx.gbl...
    > >> >> >> >>
    > >> >> >> >> I have heard this, but it only applies to ATA disks and not

    for
    > >> > Fibre
    > >> >> >> >> Channel disks in Clariion arrays. The way it was explained to
    > >> >> >> >> me
    > >> > was
    > >> >> >> >> that unlike Fibre Channel disks, ATA disks are not dual

    ported.
    > >> >> >> >> For
    > >> > a
    > >> >> >> >> Fibre Channel RAID group, each disk essentially has a

    connection
    > > to
    > >> >> > each
    > >> >> >> >> SP, so the link control card (LCC) in the disk array enclosure
    > >> >> >> >> (DAE)
    > >> >> > can
    > >> >> >> >> simply pass on commands to each individual disk. In an ATA

    DAE,
    > >> >> >> >> the
    > >> >> >> >> LCCs are actually called BCCs* (I don't recall what the "B"

    > > stands
    > >> >> >> >> for,
    > >> >> >> >> I've also heard BCCs called "paddle cards"). The BCCs incur
    > >> >> >> >> some
    > >> >> >> >> overhead when commands come from different SPs which are bound

    > > for
    > >> > the
    > >> >> >> >> same disk. It has to flip-flop back and forth and dump them

    to
    > > the
    > >> >> >> >> single ported ATA disk. This may not be technically accurate
    > >> >> >> >> but
    > >> > it's
    > >> >> >> >> the way I remember it being explained.
    > >> >> >> >>
    > >> >> >> >> The rule of thumb I've heard is that all LUNs in an ATA RAID

    > > group
    > >> >> >> >> should be owned by a single SP. For a Fibre Channel RAID

    group
    > > it
    > >> >> >> >> doesn't matter, mix and match at will. Due to

    not-so-infrequent
    > >> >> >> >> LUN
    > >> >> >> >> trespass and simply being forgetful we currently don't obey

    this
    > >> > rule
    > >> >> > to
    > >> >> >> >> the letter. Where we've mixed and matched LUN ownership in

    ATA
    > >> >> >> >> RAID
    > >> >> >> >> groups we have yet to see any problems, performance or
    > >> >> >> >> otherwise.
    > >> >> >> >>
    > >> >> >> >> * Early incarnations of Clariion ATA DAEs had many problems

    with
    > >> >> >> >> the
    > >> >> >> >> firmware on these BCC cards, which is called FRUMON code. I

    had
    > >> > more
    > >> >> >> >> than one 20 hour day thanks to these things around the FLARE

    13
    > >> > days.
    > >> >> >> >> They seem to have worked out the kinks since then, we've been
    > >> >> >> >> extremely
    > >> >> >> >> stable on ATA since about FLARE 16.
    > >> >> >> >>
    > >> >> >> >> Edwin vMierlo wrote:
    > >> >> >> >> > "he says that it is EMC's advice"
    > >> >> >> >> >
    > >> >> >> >> > well... he better back that up with a technical article from

    > > EMC
    > >> >> >> >> >
    > >> >> >> >> > or you need to talk to EMC directly, because I have never
    > >> >> >> >> > heard
    > >> >> >> >> > that,
    > >> >> >> > and I
    > >> >> >> >> > have some experience with different models in the CX-series.
    > >> >> >> >> >
    > >> >> >> >> >
    > >> >> >> >> >
    > >> >> >> >> >
    > >> >> >> >> >
    > >> >> >> >> > "Peter" wrote in message
    > >> >> >> >> > news:uK37JhRGHHA.4904@TK2MSFTNGP04.phx.gbl...
    > >> >> >> >> >> Dear Edwin,
    > >> >> >> >> >>
    > >> >> >> >> >> I hold the same opinion as yours. I have asked him and he

    > > says
    > >> >> >> >> >> that
    > >> >> >> >> >> it
    > >> >> >> > is
    > >> >> >> >> >> EMC's advice. However, I have searched PowerLink but there
    > >> >> >> >> >> is
    > >> >> >> >> >> no
    > >> >> >> > article
    > >> >> >> >> >> mentioning it.
    > >> >> >> >> >>
    > >> >> >> >> >> Besides, if I choose auto assign, the first LUN in the same

    > > RAID
    > >> >> > group
    > >> >> >> >> >> created will choose SPA, the second one choose SPB, the

    third
    > >> >> >> >> >> one
    > >> >> >> > choose
    > >> >> >> >> > SPA
    > >> >> >> >> >> that is different from what he says.
    > >> >> >> >> >>
    > >> >> >> >> >> Peter
    > >> >> >> >> >>
    > >> >> >> >> >>
    > >> >> >> >> >>
    > >> >> >> >> >> "Edwin vMierlo"
    > >> >> >> >> >> wrote
    > >> >> >> >> >> in
    > >> >> >> > message
    > >> >> >> >> >> news:%23eoD1dRGHHA.3304@TK2MSFTNGP05.phx.gbl...
    > >> >> >> >> >>> You do not have to assign them to the same SP
    > >> >> >> >> >>> I would recommend that you spread your LUNs over the two

    > > SP's,
    > >> > to
    > >> >> >> > ensure
    > >> >> >> >> >>> both SP's are processing IO
    > >> >> >> >> >>>
    > >> >> >> >> >>> being in the same RAID group has nothing to do with which

    SP
    > >> > owns
    > >> >> > the
    > >> >> >> >> > LUN
    > >> >> >> >> >>> (from experience)
    > >> >> >> >> >>>
    > >> >> >> >> >>> unless the application or configuration is specific, and

    > > there
    > >> > are
    > >> >> >> > good
    > >> >> >> >> >>> technical reasons (which we cannot see in small posts in

    the
    > >> >> >> > newsgroups)
    > >> >> >> >> >>> then there _might_ be a reason to assign all LUNs in a
    > >> >> >> >> >>> RAIDgroup
    > >> >> >> >> >>> to
    > >> >> >> > the
    > >> >> >> >> >>> same
    > >> >> >> >> >>> SP.
    > >> >> >> >> >>>
    > >> >> >> >> >>> Ask the consultant: why all luns in the same raidgroup to
    > >> >> >> >> >>> the
    > >> > same
    > >> >> > sp
    > >> >> >> > ?
    > >> >> >> >> >>> being a consultant, he/she must back up suggestions with
    > >> > technical
    > >> >> >> >> >>> reasons... (right ?)
    > >> >> >> >> >>>
    > >> >> >> >> >>> HTH,
    > >> >> >> >> >>> _Edwin.
    > >> >> >> >> >>>
    > >> >> >> >> >>> "Peter" wrote in message
    > >> >> >> >> >>> news:%23eRCESRGHHA.1468@TK2MSFTNGP04.phx.gbl...
    > >> >> >> >> >>>> Dear JM and Edwin,
    > >> >> >> >> >>>>
    > >> >> >> >> >>>> Thank you for your advice.
    > >> >> >> >> >>>>
    > >> >> >> >> >>>> A consultant suggests that for all LUNs in the same RAID

    > > Group
    > >> >> >> >> >>>> should
    > >> >> >> >> > be
    > >> >> >> >> >>>> assigned to the same SP. Is it sensible ?
    > >> >> >> >> >>>>
    > >> >> >> >> >>>> Regards,
    > >> >> >> >> >>>> Peter
    > >> >> >> >> >>>>
    > >> >> >> >> >>>> "Edwin vMierlo"

    > > wrote
    > >> > in
    > >> >> >> >> > message
    > >> >> >> >> >>>> news:OSKVPSIGHHA.1784@TK2MSFTNGP06.phx.gbl...
    > >> >> >> >> >>>>> I would agree with JM, however one small remark to

    further
    > >> >> > specify
    > >> >> >> > for
    > >> >> >> >> > a
    > >> >> >> >> >>>>> CX;
    > >> >> >> >> >>>>>
    > >> >> >> >> >>>>> if one LUN is owned by one SP and the next LUN is owned

    by
    > >> >> >> >> >>>>> the
    > >> >> >> >> >>>>> other
    > >> >> >> >> >>>>> SP,
    > >> >> >> >> >>>>> than IO will be processed by both SP's
    > >> >> >> >> >>>>> (at least that is my experience with CX)
    > >> >> >> >> >>>>>
    > >> >> >> >> >>>>> Not sure if this fully classify as an Active/Active as

    JM
    > >> >> >> >> >>>>> describes,
    > >> >> >> >> >>>>> but
    > >> >> >> >> >>>>> by
    > >> >> >> >> >>>>> spreading LUNs accross the SPs you will get the benifit

    of
    > >> > both
    > >> >> >> >> >>>>> SP's
    > >> >> >> >> >>>>> processing IO.
    > >> >> >> >> >>>>>
    > >> >> >> >> >>>>>
    > >> >> >> >> >>>>>
    > >> >> >> >> >>>>> "JM" wrote in message
    > >> >> >> >> >>>>> news:%231WXTKIGHHA.1248@TK2MSFTNGP03.phx.gbl...
    > >> >> >> >> >>>>>> Peter wrote:
    > >> >> >> >> >>>>>>> Dear JM,
    > >> >> >> >> >>>>>>>
    > >> >> >> >> >>>>>>> Many thanks for your advice.
    > >> >> >> >> >>>>>>>
    > >> >> >> >> >>>>>>> We are using EMC CX300 SAN with 2 FC and 2 SP. Is it

    an
    > >> >> >> >> >>>>>>> Active
    > >> >> > /
    > >> >> >> >> >>>>> Passive
    > >> >> >> >> >>>>>>> array ?
    > >> >> >> >> >>>>>>>
    > >> >> >> >> >>>>>>> Regards,
    > >> >> >> >> >>>>>>> Peter
    > >> >> >> >> >>>>>> By my definition, any EMC Clariion is going to be an

    > > active
    > >> >> >> >> >>>>>> /
    > >> >> >> >> > passive
    > >> >> >> >> >>>>>> array. They may call it active / active, but like I

    said
    > >> > it's
    > >> >> >> >> > really
    > >> >> >> >> >>>>>> just two SPs which are active / passive partners for

    each
    > >> >> >> >> >>>>>> other.
    > >> >> >> > If
    > >> >> >> >> >>>>>> your host has one connection to each SP, it will only

    be
    > >> >> >> >> >>>>>> able
    > >> >> >> >> >>>>>> to
    > >> >> >> >> > send
    > >> >> >> >> >>> IO
    > >> >> >> >> >>>>>> down the path to the SP which owns each individual LUN.
    > >> >> > Typically
    > >> >> >> >> > you
    > >> >> >> >> >>>>>> would install EMC's PowerPath software to help the host
    > >> >> > understand
    > >> >> >> >> > how
    > >> >> >> >> >>>>>> to manage the multiple paths back to the Clariion.
    > >> >> >> >> >>>>>
    > >> >> >> >> >>>>
    > >> >> >> >> >>>
    > >> >> >> >> >>
    > >> >> >> >> >
    > >> >> >> >> >
    > >> >> >> >
    > >> >> >> >
    > >> >> >>
    > >> >> >>
    > >> >> >
    > >> >> >
    > >> >>
    > >> >>
    > >> >
    > >> >
    > >>
    > >>

    > >
    > >

    >
    >




  2. Re: What is Active / Active Array ?


    If it makes you feel any better, we have well over 100 hosts attached to
    Clariion arrays and we don't run the Navisphere Agent on a single one of
    them. It can be helpful in registering the HBAs with the Clariion when
    initially attaching a host, and also provides useful information about
    LUN to drive letter or filesystem mapping within Navisphere. As far as
    I know, that's about all it's good for. None of our SAN attached hosts
    have any kind of IP access to the Clariion SPs. We also run PowerPath,
    which does all its communication with the Clariion over FC.

    Peter wrote:
    > Besides, I would like to know whether Navisphere Agent affects PowerPath
    > Failover software ?
    >
    > Peter
    >
    > "Edwin vMierlo" wrote in message
    > news:eh3DV8pHHHA.1468@TK2MSFTNGP04.phx.gbl...
    >> They can talk to you the security setup as well (I guess, we only have EMC
    >> kit directly from EMC, I have no explicit experience dealing with Dell in
    >> this matter)
    >>
    >> Get the connectivity guide (from EMC)... ask the Dell boys to give you a
    >> (soft-)copy, it is a handy document explaining setups in general. You can
    >> base your internal setup and maintenance procedures on this doc.
    >>
    >> good luck,
    >> Edwin.
    >>
    >>
    >> "Peter" wrote in message
    >> news:urd80xpHHHA.3780@TK2MSFTNGP02.phx.gbl...
    >>> Yes, the host can access the disks.
    >>>
    >>> Since we purchased the array from DELL, we are not able to talk to EMC
    >>> support. We have to talk to DELL staff.
    >>>
    >>> Regards,
    >>> Peter
    >>>
    >>> "Edwin vMierlo" wrote in message
    >>> news:OCcvyrpHHHA.420@TK2MSFTNGP06.phx.gbl...
    >>>> Peter,
    >>>>
    >>>> non-functioning of Navisphere Agent will not limit you accessing your
    >>>> disks,
    >>>> if your HBA's from the host are properly logged in and registered, and

    >> in
    >>>> the proper storage group, you host will be able access the disk.
    >>>>
    >>>> Q: can you access the disks from your host ?
    >>>>
    >>>> then in regards to non-functioning of the Agent, there are some
    >>>> security
    >>>> settings which need to match between host and array. Easiest is to open

    >> a
    >>>> case with EMC, they can talk you through this is 5 minutes.

    >> Alternatively,
    >>>> there is a connectivity guide on the EMC website which will detail all
    >>>> steps
    >>>> needed. I would like to talk you through it by newsgroup posts, but the
    >>>> procedure has too many variables to be 100% accurate.
    >>>>
    >>>>
    >>>>
    >>>> "Peter" wrote in message
    >>>> news:uJhuY5oHHHA.1264@TK2MSFTNGP03.phx.gbl...
    >>>>> Dear Edwin and JM,
    >>>>>
    >>>>> DELL Engineer has updated the SAN Firmware and Hosts HBA / Navisphere

    >> and
    >>>>> Powerpath for us during the weekend.
    >>>>>
    >>>>> However, from Navisphere web interface, we find that the Navisphere

    >> Agent
    >>>> in
    >>>>> one of the host (Windows 2003 Server) cannot be retrieved by the
    >>>>> Array.
    >>>>> I
    >>>>> would like to what will be the consequence ? Is it a serious problem

    >> or
    >>>>> just the drives information is not presented ?
    >>>>>
    >>>>> I have checked Powerpath and it seems to be OK.
    >>>>>
    >>>>> Regards,
    >>>>> Peter
    >>>>> "Edwin vMierlo" wrote in

    >> message
    >>>>> news:ONtpYYeGHHA.3952@TK2MSFTNGP02.phx.gbl...
    >>>>>> response in-line
    >>>>>>
    >>>>>>
    >>>>>> "Peter" wrote in message
    >>>>>> news:ObW4iHeGHHA.924@TK2MSFTNGP02.phx.gbl...
    >>>>>>> Dear JM and Edwin,
    >>>>>>>
    >>>>>>> Thank you for your advice.
    >>>>>>> Since we are using FC disk and not ATA one, the consultant's advice
    >>>> looks
    >>>>>>> not valid.
    >>>>>>>
    >>>>>>> Does ATA disk means SATA disk ?
    >>>>>>> Can we use FC disk with different speed ?
    >>>>>> This might be depending on the Flare-code revision, and other
    >>>>>> things,
    >>>> you
    >>>>>> really need to ask that question to your EMC rep.
    >>>>>>
    >>>>>>> This is because from EMC web
    >>>>>>> site, I find that they support SAN disk with higher rpm (Of course,
    >>>> 300GB
    >>>>>> HD
    >>>>>>> is still not available).
    >>>>>>>
    >>>>>>> The consultant also suggests us to upgrade to the latest firmware

    >> (The
    >>>>>>> one
    >>>>>>> we are using is more than 2 years old). Is this suggestion valid ?
    >>>>>> Most of the time; yes. Keep abrest with latest versions of software

    >> and
    >>>>>> firmware. BUT If you have a test or try-out system, I would highly
    >>>>>> recommend
    >>>>>> to run upgrades on test systems first.
    >>>>>>
    >>>>>>> Regards,
    >>>>>>> Peter
    >>>>>>>
    >>>>>>> "Edwin vMierlo" wrote in
    >>>> message
    >>>>>>> news:eup4p1cGHHA.1240@TK2MSFTNGP03.phx.gbl...
    >>>>>>>> That could explain it, I have no experience with CX series with

    >> ATA
    >>>>>>>> drives,
    >>>>>>>> only FC drives.
    >>>>>>>>
    >>>>>>>> thanks for explaingin JM !
    >>>>>>>> rgds,
    >>>>>>>> Edwin.
    >>>>>>>>
    >>>>>>>>
    >>>>>>>> "JM" wrote in message
    >>>>>>>> news:%23jOAPHVGHHA.1280@TK2MSFTNGP04.phx.gbl...
    >>>>>>>>> I have heard this, but it only applies to ATA disks and not for
    >>>> Fibre
    >>>>>>>>> Channel disks in Clariion arrays. The way it was explained to
    >>>>>>>>> me
    >>>> was
    >>>>>>>>> that unlike Fibre Channel disks, ATA disks are not dual ported.
    >>>>>>>>> For
    >>>> a
    >>>>>>>>> Fibre Channel RAID group, each disk essentially has a connection

    >> to
    >>>>>> each
    >>>>>>>>> SP, so the link control card (LCC) in the disk array enclosure
    >>>>>>>>> (DAE)
    >>>>>> can
    >>>>>>>>> simply pass on commands to each individual disk. In an ATA DAE,
    >>>>>>>>> the
    >>>>>>>>> LCCs are actually called BCCs* (I don't recall what the "B"

    >> stands
    >>>>>>>>> for,
    >>>>>>>>> I've also heard BCCs called "paddle cards"). The BCCs incur
    >>>>>>>>> some
    >>>>>>>>> overhead when commands come from different SPs which are bound

    >> for
    >>>> the
    >>>>>>>>> same disk. It has to flip-flop back and forth and dump them to

    >> the
    >>>>>>>>> single ported ATA disk. This may not be technically accurate
    >>>>>>>>> but
    >>>> it's
    >>>>>>>>> the way I remember it being explained.
    >>>>>>>>>
    >>>>>>>>> The rule of thumb I've heard is that all LUNs in an ATA RAID

    >> group
    >>>>>>>>> should be owned by a single SP. For a Fibre Channel RAID group

    >> it
    >>>>>>>>> doesn't matter, mix and match at will. Due to not-so-infrequent
    >>>>>>>>> LUN
    >>>>>>>>> trespass and simply being forgetful we currently don't obey this
    >>>> rule
    >>>>>> to
    >>>>>>>>> the letter. Where we've mixed and matched LUN ownership in ATA
    >>>>>>>>> RAID
    >>>>>>>>> groups we have yet to see any problems, performance or
    >>>>>>>>> otherwise.
    >>>>>>>>>
    >>>>>>>>> * Early incarnations of Clariion ATA DAEs had many problems with
    >>>>>>>>> the
    >>>>>>>>> firmware on these BCC cards, which is called FRUMON code. I had
    >>>> more
    >>>>>>>>> than one 20 hour day thanks to these things around the FLARE 13
    >>>> days.
    >>>>>>>>> They seem to have worked out the kinks since then, we've been
    >>>>>>>>> extremely
    >>>>>>>>> stable on ATA since about FLARE 16.
    >>>>>>>>>
    >>>>>>>>> Edwin vMierlo wrote:
    >>>>>>>>>> "he says that it is EMC's advice"
    >>>>>>>>>>
    >>>>>>>>>> well... he better back that up with a technical article from

    >> EMC
    >>>>>>>>>> or you need to talk to EMC directly, because I have never
    >>>>>>>>>> heard
    >>>>>>>>>> that,
    >>>>>>>> and I
    >>>>>>>>>> have some experience with different models in the CX-series.
    >>>>>>>>>>
    >>>>>>>>>>
    >>>>>>>>>>
    >>>>>>>>>>
    >>>>>>>>>>
    >>>>>>>>>> "Peter" wrote in message
    >>>>>>>>>> news:uK37JhRGHHA.4904@TK2MSFTNGP04.phx.gbl...
    >>>>>>>>>>> Dear Edwin,
    >>>>>>>>>>>
    >>>>>>>>>>> I hold the same opinion as yours. I have asked him and he

    >> says
    >>>>>>>>>>> that
    >>>>>>>>>>> it
    >>>>>>>> is
    >>>>>>>>>>> EMC's advice. However, I have searched PowerLink but there
    >>>>>>>>>>> is
    >>>>>>>>>>> no
    >>>>>>>> article
    >>>>>>>>>>> mentioning it.
    >>>>>>>>>>>
    >>>>>>>>>>> Besides, if I choose auto assign, the first LUN in the same

    >> RAID
    >>>>>> group
    >>>>>>>>>>> created will choose SPA, the second one choose SPB, the third
    >>>>>>>>>>> one
    >>>>>>>> choose
    >>>>>>>>>> SPA
    >>>>>>>>>>> that is different from what he says.
    >>>>>>>>>>>
    >>>>>>>>>>> Peter
    >>>>>>>>>>>
    >>>>>>>>>>>
    >>>>>>>>>>>
    >>>>>>>>>>> "Edwin vMierlo"
    >>>>>>>>>>> wrote
    >>>>>>>>>>> in
    >>>>>>>> message
    >>>>>>>>>>> news:%23eoD1dRGHHA.3304@TK2MSFTNGP05.phx.gbl...
    >>>>>>>>>>>> You do not have to assign them to the same SP
    >>>>>>>>>>>> I would recommend that you spread your LUNs over the two

    >> SP's,
    >>>> to
    >>>>>>>> ensure
    >>>>>>>>>>>> both SP's are processing IO
    >>>>>>>>>>>>
    >>>>>>>>>>>> being in the same RAID group has nothing to do with which SP
    >>>> owns
    >>>>>> the
    >>>>>>>>>> LUN
    >>>>>>>>>>>> (from experience)
    >>>>>>>>>>>>
    >>>>>>>>>>>> unless the application or configuration is specific, and

    >> there
    >>>> are
    >>>>>>>> good
    >>>>>>>>>>>> technical reasons (which we cannot see in small posts in the
    >>>>>>>> newsgroups)
    >>>>>>>>>>>> then there _might_ be a reason to assign all LUNs in a
    >>>>>>>>>>>> RAIDgroup
    >>>>>>>>>>>> to
    >>>>>>>> the
    >>>>>>>>>>>> same
    >>>>>>>>>>>> SP.
    >>>>>>>>>>>>
    >>>>>>>>>>>> Ask the consultant: why all luns in the same raidgroup to
    >>>>>>>>>>>> the
    >>>> same
    >>>>>> sp
    >>>>>>>> ?
    >>>>>>>>>>>> being a consultant, he/she must back up suggestions with
    >>>> technical
    >>>>>>>>>>>> reasons... (right ?)
    >>>>>>>>>>>>
    >>>>>>>>>>>> HTH,
    >>>>>>>>>>>> _Edwin.
    >>>>>>>>>>>>
    >>>>>>>>>>>> "Peter" wrote in message
    >>>>>>>>>>>> news:%23eRCESRGHHA.1468@TK2MSFTNGP04.phx.gbl...
    >>>>>>>>>>>>> Dear JM and Edwin,
    >>>>>>>>>>>>>
    >>>>>>>>>>>>> Thank you for your advice.
    >>>>>>>>>>>>>
    >>>>>>>>>>>>> A consultant suggests that for all LUNs in the same RAID

    >> Group
    >>>>>>>>>>>>> should
    >>>>>>>>>> be
    >>>>>>>>>>>>> assigned to the same SP. Is it sensible ?
    >>>>>>>>>>>>>
    >>>>>>>>>>>>> Regards,
    >>>>>>>>>>>>> Peter
    >>>>>>>>>>>>>
    >>>>>>>>>>>>> "Edwin vMierlo"

    >> wrote
    >>>> in
    >>>>>>>>>> message
    >>>>>>>>>>>>> news:OSKVPSIGHHA.1784@TK2MSFTNGP06.phx.gbl...
    >>>>>>>>>>>>>> I would agree with JM, however one small remark to further
    >>>>>> specify
    >>>>>>>> for
    >>>>>>>>>> a
    >>>>>>>>>>>>>> CX;
    >>>>>>>>>>>>>>
    >>>>>>>>>>>>>> if one LUN is owned by one SP and the next LUN is owned by
    >>>>>>>>>>>>>> the
    >>>>>>>>>>>>>> other
    >>>>>>>>>>>>>> SP,
    >>>>>>>>>>>>>> than IO will be processed by both SP's
    >>>>>>>>>>>>>> (at least that is my experience with CX)
    >>>>>>>>>>>>>>
    >>>>>>>>>>>>>> Not sure if this fully classify as an Active/Active as JM
    >>>>>>>>>>>>>> describes,
    >>>>>>>>>>>>>> but
    >>>>>>>>>>>>>> by
    >>>>>>>>>>>>>> spreading LUNs accross the SPs you will get the benifit of
    >>>> both
    >>>>>>>>>>>>>> SP's
    >>>>>>>>>>>>>> processing IO.
    >>>>>>>>>>>>>>
    >>>>>>>>>>>>>>
    >>>>>>>>>>>>>>
    >>>>>>>>>>>>>> "JM" wrote in message
    >>>>>>>>>>>>>> news:%231WXTKIGHHA.1248@TK2MSFTNGP03.phx.gbl...
    >>>>>>>>>>>>>>> Peter wrote:
    >>>>>>>>>>>>>>>> Dear JM,
    >>>>>>>>>>>>>>>>
    >>>>>>>>>>>>>>>> Many thanks for your advice.
    >>>>>>>>>>>>>>>>
    >>>>>>>>>>>>>>>> We are using EMC CX300 SAN with 2 FC and 2 SP. Is it an
    >>>>>>>>>>>>>>>> Active
    >>>>>> /
    >>>>>>>>>>>>>> Passive
    >>>>>>>>>>>>>>>> array ?
    >>>>>>>>>>>>>>>>
    >>>>>>>>>>>>>>>> Regards,
    >>>>>>>>>>>>>>>> Peter
    >>>>>>>>>>>>>>> By my definition, any EMC Clariion is going to be an

    >> active
    >>>>>>>>>>>>>>> /
    >>>>>>>>>> passive
    >>>>>>>>>>>>>>> array. They may call it active / active, but like I said
    >>>> it's
    >>>>>>>>>> really
    >>>>>>>>>>>>>>> just two SPs which are active / passive partners for each
    >>>>>>>>>>>>>>> other.
    >>>>>>>> If
    >>>>>>>>>>>>>>> your host has one connection to each SP, it will only be
    >>>>>>>>>>>>>>> able
    >>>>>>>>>>>>>>> to
    >>>>>>>>>> send
    >>>>>>>>>>>> IO
    >>>>>>>>>>>>>>> down the path to the SP which owns each individual LUN.
    >>>>>> Typically
    >>>>>>>>>> you
    >>>>>>>>>>>>>>> would install EMC's PowerPath software to help the host
    >>>>>> understand
    >>>>>>>>>> how
    >>>>>>>>>>>>>>> to manage the multiple paths back to the Clariion.
    >>>>>>>>>>
    >>>>>>>>
    >>>>>>>
    >>>>>>
    >>>>>
    >>>>
    >>>

    >>

    >
    >


+ Reply to Thread
Page 2 of 2 FirstFirst 1 2